feat: expose fill-color ink classification on ImageInfo (#1688)

src/ocrmypdf/pdfinfo/_image.py

@@ -36,6 +36,7 @@ from ocrmypdf.pdfinfo._types import (
     UNIT_SQUARE,
     Colorspace,
     Encoding,
+    Ink,
 )
 
 logger = logging.getLogger()
@@ -61,10 +62,12 @@ class ImageInfo:
         pdfimage: Object | None = None,
         inline: PdfInlineImage | None = None,
         shorthand=None,
+        fill_ink: Ink | None = None,
     ):
         """Initialize an ImageInfo."""
         self._name = str(name)
         self._shorthand = shorthand
+        self._fill_ink = fill_ink
 
         pim: PdfInlineImage | PdfImage
 
@@ -175,6 +178,17 @@ class ImageInfo:
         """Type of image, either 'image' or 'stencil'."""
         return self._type
 
+    @property
+    def ink(self) -> Ink | None:
+        """Fill-color classification for stencil masks, else None.
+
+        A stencil (image mask) is painted with the current fill color; this
+        reports whether that color is mono/gray/color so the rasterizer can
+        choose a device that does not discard the distinction. Non-stencil
+        images return None.
+        """
+        return self._fill_ink if self._type == 'stencil' else None
+
     @property
     def width(self) -> int:
         """Width of the image in pixels."""
@@ -249,7 +263,10 @@ def _find_inline_images(contentsinfo: ContentsInfo) -> Iterator[ImageInfo]:
     """Find inline images in the contentstream."""
     for n, inline in enumerate(contentsinfo.inline_images):
         yield ImageInfo(
-            name=f'inline-{n:02d}', shorthand=inline.shorthand, inline=inline.iimage
+            name=f'inline-{n:02d}',
+            shorthand=inline.shorthand,
+            inline=inline.iimage,
+            fill_ink=inline.fill_ink,
         )
 
 
@@ -300,7 +317,12 @@ def _find_regular_images(
                 # these from our DPI calculation for the page.
                 continue
 
-            yield ImageInfo(name=draw.name, pdfimage=pdfimage, shorthand=draw.shorthand)
+            yield ImageInfo(
+                name=draw.name,
+                pdfimage=pdfimage,
+                shorthand=draw.shorthand,
+                fill_ink=draw.fill_ink,
+            )
 
 
 def _find_form_xobject_images(pdf: Pdf, container: Object, contentsinfo: ContentsInfo):

tests/test_pdfinfo.py

@@ -371,3 +371,64 @@ def test_fill_ink_tolerates_malformed_color_operands(body):
 )
 def test_ink_from_components(space, comps, expected):
     assert _ink_from_components(space, comps) is Ink[expected]
+
+
+def _make_image_mask_pdf(path, content_fill: bytes):
+    """Build a 1-page PDF with one 8x8 image mask painted with content_fill.
+
+    content_fill is the color operator sequence emitted before drawing the
+    mask, e.g. b"0.263 0.263 0.263 rg".
+    """
+    pdf = pikepdf.Pdf.new()
+    pdf.add_blank_page(page_size=(72, 72))
+    # 8x8 1-bpc mask, each row padded to a byte (1 byte per row).
+    mask_bytes = bytes([0x7E] * 8)
+    mask = pikepdf.Stream(pdf, mask_bytes)
+    mask.Type = pikepdf.Name.XObject
+    mask.Subtype = pikepdf.Name.Image
+    mask.Width = 8
+    mask.Height = 8
+    mask.ImageMask = True
+    mask.BitsPerComponent = 1
+    name = pdf.pages[0].add_resource(mask, pikepdf.Name.XObject)
+    pdf.pages[0].Contents = pikepdf.Stream(
+        pdf, b"q 72 0 0 72 0 0 cm %s %s Do Q" % (content_fill, bytes(name))
+    )
+    pdf.save(path)
+    return path
+
+
+@pytest.fixture
+def mask_gray_pdf(outdir):
+    return _make_image_mask_pdf(outdir / 'mask_gray.pdf', b"0.263 0.263 0.263 rg")
+
+
+@pytest.fixture
+def mask_rgb_pdf(outdir):
+    return _make_image_mask_pdf(outdir / 'mask_rgb.pdf', b"0.8 0.2 0.2 rg")
+
+
+@pytest.fixture
+def mask_black_pdf(outdir):
+    return _make_image_mask_pdf(outdir / 'mask_black.pdf', b"0 g")
+
+
+def test_imageinfo_ink_gray(mask_gray_pdf):
+    image = pdfinfo.PdfInfo(mask_gray_pdf)[0].images[0]
+    assert image.type_ == 'stencil'
+    assert image.ink is Ink.gray
+
+
+def test_imageinfo_ink_color(mask_rgb_pdf):
+    image = pdfinfo.PdfInfo(mask_rgb_pdf)[0].images[0]
+    assert image.ink is Ink.color
+
+
+def test_imageinfo_ink_black(mask_black_pdf):
+    image = pdfinfo.PdfInfo(mask_black_pdf)[0].images[0]
+    assert image.ink is Ink.mono
+
+
+def test_imageinfo_ink_none_for_regular_image(eight_by_eight_regular_image):
+    image = pdfinfo.PdfInfo(eight_by_eight_regular_image)[0].images[0]
+    assert image.ink is None